One doc tagged with "ai-security"

Hackers use CSS and PDF hacks to hide malicious prompts from humans while keeping them visible to LLMs. Learn how to detect stealth attacks.